EP2412117B1

EP2412117B1 - Adaptive precoding codebooks for wireless communications

Info

Publication number: EP2412117B1
Application number: EP10755427.1A
Authority: EP
Inventors: Pengfei Xia; Tang Yang; Zhigang Rong
Original assignee: Huawei Technologies Co Ltd
Current assignee: Huawei Technologies Co Ltd
Priority date: 2009-03-23
Filing date: 2010-03-23
Publication date: 2013-05-01
Anticipated expiration: 2030-03-23
Also published as: CN102308508A; US20100238913A1; US8675627B2; EP2412117A4; CN102308508B; WO2010108427A1; EP2412117A1

Description

TECHNICAL FIELD

The present invention relates generally to wireless communications, and more particularly to adaptive precoding codebooks for wireless communications.

BACKGROUND

In a wireless communications system, the communications system's capacity may be significantly improved when a transmitter (also referred to as a base station) has full or partial knowledge of a channel over which it will be transmitting. Information related to the channel may be referred to as channel state information (CSI). CSI may be obtained by the transmitter over a reverse feedback link through the channel. A receiver (also referred to as a mobile user end or mobile station or user equipment) of the transmissions made by the transmitter may transmit CSI back to the transmitter over the reverse feedback link through the channel.
In MIMO (multiple input multiple output) wireless communications, channel state information can be estimated at the wireless receiver and fed back to the wireless transmitter, so that the transmitter can adapt its transmit scheme to the wireless channel. Such feedback is usually achieved through a codebook, which serves as a common dictionary of the current channel condition to the wireless transmitter and receiver.
However, in practice, changes in the wireless channel can be much more severe than what can be accommodated by a static codebook. Using a dynamic codebook, as e.g. disclosed in WO 2007/050861 , overcomes this problem but introduces the problem of increased complexity.

SUMMARY OF THE INVENTION

These and other problems are generally solved or circumvented, and technical advantages are generally achieved, by illustrative embodiments of the present invention.
The following detailed description is not to be taken in a limiting sense, and the scope of the present invention is defined only by the appended claims.
In accordance with an aspect of the invention of the present invention, a method of wireless communication comprises reading a baseline codebook of at least a rank-2, the baseline codebook comprising codewords representing precoding matrices. The method further comprises generating an adaptive codebook by multiplying a first column of the codeword with a first transform matrix calculated from a channel correlation matrix, and multiplying a second column of the codeword with a second transform matrix calculated from the channel correlation matrix. The first transform matrix is orthogonal to the second transform matrix.
The foregoing has outlined rather broadly the features of an aspect of the invention of the present invention in order that the detailed description of the invention that follows may be better understood. Additional features and advantages of aspect of the inventions of the invention will be described hereinafter, which form the subject of the claims of the invention. It should be appreciated by those skilled in the art that the conception and specific aspect of the inventions disclosed may be readily utilized as a basis for modifying or designing other structures or processes for carrying out the same purposes of the present invention. It should also be realized by those skilled in the art that such equivalent constructions do not depart from the invention as set forth in the appended claims.

BRIEF DESCRIPTION OF THE DRAWINGS

For a more complete understanding of the present invention, and the advantages thereof, reference is now made to the following descriptions taken in conjunction with the accompanying drawings, in which:

Figure 1 is a portion of a wireless communication system, in accordance with an embodiment of the invention;
Figure 2 illustrates a flow diagram of a receiver in operation in accordance with an embodiment of the invention; and
Figure 3 illustrates a flow diagram of a transmitter in operation in accordance with an embodiment of the invention.

Corresponding numerals and symbols in the different figures generally refer to corresponding parts unless otherwise indicated. The figures are drawn to clearly illustrate the relevant aspects of the embodiments and are not necessarily drawn to scale.

DETAILED DESCRIPTION OF ILLUSTRATIVE EMBODIMENTS

The making and using of various embodiments are discussed in detail below. It should be appreciated, however, that the present invention provides many applicable inventive concepts that can be embodied in a wide variety of specific contexts. The specific embodiments discussed are merely illustrative of specific ways to make and use the invention, and do not limit the scope of the invention.
Embodiments of the invention enable significant performance improvements by the use of dynamic precoding codebooks that incorporate spatial channel correlation data to a baseline codebook. In various embodiments, the transformations from the baseline codebook are carefully selected to minimize complexity while preserving a unitary codebook.
Using Figure 1, a wireless communications system will be described in accordance with an embodiment of the invention. Flow diagrams at the receiver unit 110 and transmitter unit 105 are described in accordance with embodiments of the invention in Figures 2 and 3.
Figure 1 is a portion of a wireless communication system, in accordance with an embodiment of the invention. A compressed digital source in the form of a binary data stream 2 is fed to a simplified transmitting block 5 within the transmitter unit 105, which is a base station or the like, or a router, e.g., on a home MIMO network. The transmitting block 5 encompasses the functions of error control coding and (possibly joined with) mapping to complex modulation symbols. The simplified transmitting block 5 produces several separate symbol streams which range from independent to partially redundant to fully redundant. Each of the symbol streams is mapped to one of the multiple transmitter antennas 12-18. After upward frequency conversion, filtering and amplification, the transmit signals 6 are transmitted to the transmitter 7 and launched into the wireless channel 115 by the multiple transmitter antennas 12-18 (four antennas are illustrated only as an example).
The wireless transmitter unit 105 communicates with the receiver unit 110 via the wireless channel 115. As an illustration, the wireless transmitter unit 105 comprises four transmit antennas 12, 14, 16, and 18 and the receiver unit 110 comprises four receive antennas 22, 24, 26, and 28. In other embodiments, any number of transmit antennas and any number of receive antennas are used to form a link through the channel 115. The transmitter unit 105 dynamically tailors the transmit signals 6 to be transmitted into the channel 115 in a manner, for example, that improves channel throughput or lowers bit error rate or both.
The signals transmitted from the transmit antennas 12, 14, 16, and 18 are received at the four receive antennas 22, 24, 26, and 28 of the receiver 21. Subsequently, demodulation and demapping operations are performed in demapper/demodulator 25 of the receiver unit 110 to recover the message. The level of intelligence, complexity, and a priori channel knowledge used in selecting the coding and antenna mapping algorithms are adapted during the link depending on the application and the nature of the transmission.
The receiver unit 110 transmits channel-related feedback information (or channel state information) to the transmitter unit 105 for use by the transmitting block 5 in developing transmit signals 6. The receiver unit 110 generates the feedback information by, for example, appropriately processing reference information received from the transmitter unit 105.
While providing detailed feedback of the channel will help to increase the spectral efficiency of the channel, practical limitations frequently prohibit its practical application. For example, detailed feedback can take up precious bandwidth and processing power. Further, the channel may vary over time and hence the feedback obtained at the transmitter is prone to imperfections such as feedback delay, feedback error, channel estimation error, and the like. Hence, a limited feedback is more preferred. On the other hand, non-adaptive designs using no channel state information suffer from poor performance.
As a consequence, transmit beamforming with limited feedback provides significant performance gains. To minimize information needed to provide feedback, a codebook is used. The beamforming codebook is designed and maintained at both transmitter unit 105 and the receiver unit 110 to facilitate the operation of limited feedback transmit beamforming. The codebook can be a collection of candidate precoding matrices/vectors to be used in generating the transmit signals 6. Codebooks are typically designed to match a particular type of channel characteristics, for example. For single user multiple input multiple output (SU-MIMO) independent and identically distributed (i.i.d.) Rayleigh fading channels, Grassmannian line/subspace packing (GLP) based codebook can achieve near optimal performance. On the other hand, these GLP codebooks do not perform well under spatially correlated fading channels.
Further, the above codebook designs are fixed in the sense that the codebook never changes during the link. As a consequence, while preferable for simplicity, the codebook does not perform well in drastically different fading scenarios (e.g., suburban macro fading, urban micro fading, urban macro fading), for different antenna spacing (e.g., 0.5□, 4□), different mobile speed, different antenna pattern and polarization profile.
Spatial correlation assisted codebook design is one way of improving system performance further. In this design, a fixed baseline codebook is first chosen from conventional designs (e.g., GLP, DFT, Householder or others). Then, the actual codebook is derived from the fixed baseline codebook depending on the channel. For example, the transformation used to derive the actual codebook may be based on medium to long term statistics of the channel (e.g., spatial channel correlation). The statistical channel information may be obtained at the transmitter either from infrequent feedback, or from uplink sounding signals assuming channel reciprocity at the channel statistics level. Thus, a varying codebook design is implemented that adapts the codebook to match the underlying channel characteristics.
Embodiments of the invention will now be described mathematically. First, a rank-1 adaptive codebook is described followed by embodiments of the invention describing a rank-2 adaptive codebook.
The received signal y at the receiver unit 110 is represented as: $y = H θ s + n,$
In any MIMO, there are Nt antennas at the transmitter unit 105 and Nr antennas at a receiver unit 110. Hence, there are Nt x Nr MIMO channels between the transmitter unit 105 and the receiver unit 110. In equation 1, H is the channel matrix of size Nr x Nt, Θ is a unit norm beamformer of size Nt x 1 selected from a pre-defined codebook W1d , s is a scalar information symbol, and n is the additive white Gaussian noise.
The pre-defined codebook W1d = {w1, ... ,wN} is a collection of N unit-norm candidate beamforming vectors each of size Nt x 1, with number of feedback bits B=log2N. Without loss of generality, let H be a zero mean and R=E[H'H] be the transmitter side channel correlation matrix. The eigenvalue decomposition of the channel correlation matrix is given as: $R = \sum_{i = 1}^{N_{t}} σ_{i}_{2} v_{i} v_{i}^{ʹ} .$
Here σ12 ,... ,σNt2 are the eigenvalues in a decreasing order and V=[v1,...,vNt] is the unitary matrix collecting the eigenvectors in a corresponding order. A reduced rank representation of transmitter side channel correlation matrix R is defined as $R_{ℓ, k} = \sum_{i = ℓ}^{k} σ_{i}_{2} v_{i} v_{i}^{ʹ},$

Equation 3 captures channel correlation matrix R from the 1th eigenmode to the kth eigenmode. If the fixed baseline codebook is W1d, then the adaptive codebook $W_{full}^{1 d}$
is obtained by transformation as: $W_{full}^{1 d} = N (R W_{base}^{1 d}),$
In equation 4, the function N() performs column-wise vector normalization to maintain that each codeword in the adaptive codebook is a unit-norm vector. In other words, the new codebook is simply a transformed version of the baseline codebook W1d, while the transformation is determined by the underlying statistical channel information (transmitter side channel correlation matrix R). As a consequence, the adapted codebook $W_{full}^{1 d}$
has exactly the same dimension as the baseline codebook W1d. The subscript "full" indicates the fact that a full representation of transmitter side channel correlation matrix R is used to transform the baseline codebook.
Another possible transformation is to use the reduced rank representation as: $W_{reduced}^{1 d} = N (R_{1, k} W_{base}^{1 d}),$

where only the first (strongest) k eigenmodes of the transmitter side channel correlation matrix R are used in the transformation. This reduces feedback overhead associated with using the full channel correlation matrix R. The subscript "reduced" indicates that only a reduced representation of the channel correlation matrix R is used to transform the baseline codebook.
Yet another possible transformation is to use $W_{reduced}^{1 d} = N (R_{1, k}^{m} W_{base}^{1 d}),$

where m is a power factor that may be different than 1. For most applications, m=1/2 (square root) and 1 perform similarly. In various embodiments described herein, a value of m=1 is used for consistency, although in other embodiments different m's are possible.
Embodiments of the invention describe transforming a higher rank codebook using statistical channel information. As an illustration, a dual and a three layer adaptive codebook are described although the technique can be generalized for higher rank codebooks in various embodiments.
The received signal of a rank-2 SU-MIMO is defined as $y = HΘs + n,$

wherein H is the channel matrix of size N_r x N_t quantized by a pre-defined codebook F^2d, Θ is a unitary precoder of size N_t x 2, s is the 2 x 1 information symbol vector, and n is the N_r x 1 2d additive white Gaussian noise. Here, F is a collection of N unit-norm candidate codewords, with each codeword being an N_t x 2 unitary matrix. If the channel matrix H comprises a mean of zero, the channel correlation matrix R may be represented as R=E[H'H]. In particular, let the k ^th codeword of the rank-2 baseline codebooks is defined as: ${\{F_{base}^{}\}}_{k} = [\begin{matrix} f_{k 1}, & f_{k 2} \end{matrix}],$

where f _k1 ,f _k2 are the two columns of the k ^th codeword which is a rank-2 unitary matrix. Different transformation methods may be used in adapting such a rank-2 codeword by taking advantage of the channel statistics. First, the transformation used in the rank-1 case, as defined in (4), can be generalized directly as follows ${\{F_{same}^{2 d}\}}_{k} = GRAM ([\begin{matrix} R_{1, N_{t}} f_{k 1}, & R_{1, N_{t}} f_{k 2} \end{matrix}]),$

i.e., both columns of the codeword are transformed by the same channel correlation matrix R _1,Nt and hence represented with the subscript "same." The function GRAM() performs the standard Gram-Schmidt orthonormalization process for a general non-unitary matrix such that the resultant codeword is guaranteed to be unitary.
Another design uses a different correlation matrix to transform each column of the codeword. For example, in ${\{F_{diff}^{}\}}_{k} = GRAM ([\begin{matrix} R_{1, N_{t}} f_{k 1}, & R_{2, N_{t}} f_{k 2} \end{matrix}]),$

wherein the first column is transformed by the original channel correlation matrix R_1,Nt, while the second column is transformed by R_2,Nt, which is the channel correlation matrix excluding the strongest eigenmode. Hence, the transformed codebook is represented with the subscript "diff." Advantageously, equation 10 converges faster as the first column would converge to v₁ after transformation by R_1,Nt and that the second column would converge to v₂ after transformation by R_2,Nt. This method is disclosed in Provisional Application No. 61/166,133 , "System and Method for Precoding Codebook Adaptation Using Statistical Channel Information," filed on April 2, 2009, which is incorporated herein by reference. Nevertheless, the Gram-Schmidt orthonormalization process carried out afterwards damages the convergence, since R_1,Nt f_k1 and R_2,Nt f_k2 in general are not orthogonal to each other.
Embodiments of the invention overcome these problems by providing a scheme for fast convergence without doing a Gram-Schmidt orthonormalization process over the whole matrix. In accordance with an embodiment of the invention, the kth codeword of the baseline codebook F2d is transformed as ${\{F_{oddeυen}^{2 d}\}}_{k} = [N \begin{matrix} (R_{odd} f_{k 1}), & N (\begin{matrix} R_{eυen} f_{k 2} \end{matrix}) \end{matrix}] .$

where R_odd captures the joint contribution from the odd eigenmodes of the channel correlation matrix R and defined as: $R_{odd} = \sum_{i = 1, 3, 5, \dots} σ_{i}^{2} v_{i} v_{i}^{ʹ},$

and R_even captures the joint contribution from the even eigenmodes of channel correlation matrix R and defined as: $R_{eυen} = \sum_{i = 2, 4, 6, \dots} σ_{i}^{2} v_{i} v_{i}^{ʹ},$
Hence, the derived codebook is represented with subscript of "oddeven." Importantly, Rodd fk1 is orthogonal to Revenfk2 by construction, for arbitrary initial vectors of fk1, fk2. Consequently, and advantageously, no Gram-Schmidt orthonormalization is required. The odd-even adaptive codebook $F_{oddeven}$ $_{2 d}$
is thus designed as: $⌊ {\{F_{oddenven}^{}\}}_{1}, {\{F_{oddenven}^{}\}}_{2}, {\{F_{oddenven}^{}\}}_{3}, {\{F_{oddenven}^{}\}}_{k}, \dots, {\{F_{oddenven}^{}\}}_{N} ⌋$
In various embodiments, a transformed codebook is obtained for higher rank codebooks by adapting a baseline codebook with spatial channel information. In one embodiment, the method uses a first transformation matrix to transform the first column of each codeword within the codebook, and uses a different second transformation matrix to transform the second column of each codeword within the codebook. The transformation matrices are carefully chosen such that they are orthogonal so that no Gram-Schmidt orthonormalization or similar procedure is needed to make the transformed codeword a valid unitary transmit precoding matrix.
In one embodiment, for designing a rank-2 adaptive codebook, odd eigenmodes (eigenmode 1, 3, 5...) of the channel correlation matrix R transform the first column of the codeword, while even eigenmodes (eigenmode 2, 4, 6...) of the channel correlation matrix R transform the second column of the codeword.
An alternative embodiment of the invention for a rank-3 codebook is now described. In one embodiment, a rank-3 adaptive codebook (F3d) is designed using a first set of eigenmodes (eigenmodes 1, 4, 7,...) of the channel correlation matrix R to transform the first column of the codeword. A second set of eigenmodes ( eigenmodes 2, 5, 8,...) of the channel correlation matrix R transforms the second column of the baseline codeword. Finally, a third set of eigenmodes (eigenmodes 3, 6, 9,...) of the channel correlation matrix R transform the third column of the baseline codeword. The kth codeword of the rank-3 adaptive codebook (F3d) is thus designed as: ${\{F^{3 d}\}}_{k} = [N (R_{first} f_{k 1}), N (R_{second} f_{k 2}), N (R_{third} f_{k 3})]$
In various embodiments, the first transformation matrix A for the first column and the second transformation matrix B are selected to be orthogonal such that AB' =0. B' is the transpose of the second transformation matrix B. The first and the second transformation matrices A and B are functions of the channel correlation matrix R. The adaptive code book is then obtained using the transformation defined as: ${\{F_{ortho}^{}\}}_{k} = [A f_{k 1}, B f_{k 2}]$
Because of the orthogonal transformation (Afk1 and Bfk2), both the transformed columns Afk1 and Bfk2 remain orthogonal even after the transformation. As a consequence, a Gram-Schmidt orthogonalization of the whole codeword $(F_{ortho})$ $(_{2 d})$
is avoided, which is otherwise required to form valid unitary matrices.
In various embodiments, the selection of odd eigenmodes (eigenmode 1, 3, 5...) of the channel correlation matrix R as the first transformation matrix A and the even eigenmodes ( eigenmode 2, 4, 6... ) as the second transformation matrix B is one way of ensuring that the first and the second transformation matrices A and B are orthogonal.
An embodiment of the invention for a rank-3 codebook is now described. In various embodiments, a rank-3 adaptive codebook $F_{ortho}^{3 d}$
is designed by extending the prior embodiment for a rank-2 adaptive codebook. The first transformation matrix A for the first column and the second transformation matrix B are selected to be orthogonal. Similarly, a third transformation matrix C is chosen to be orthogonal to the first and the second transformation matrices. The first, the second, and the third transformation matrices A, B, and C are functions of the channel correlation matrix R. The adaptive code book is obtained using the transformation defined as: ${\{F_{ortho}^{}\}}_{k} = [A f_{k 1}, B f_{k 2}, C f_{k 3}]$
Knowledge of the spatial channel correlation matrix is needed at both the transmitter and receiver to perform the transformations at both the transmitter unit 105 and the receiver unit 110. Assuming that channel correlation is available to both transmitter unit 105 and receiver unit 110, not much additional information is needed to enable the adaptive codebooks. A signaling bit may be used to synchronize the transmitter unit 105 and receiver unit 110 to ensure that they use the common codebook, which may either be the baseline codebook or a transformed codebook based on common knowledge of the channel correlation matrix.
Using embodiments of the invention, the codebook adapts to spatial channel correlations achieving significant performance gains. Because of the adaptive nature of the codebook, system performance is less dependent on the particular choice of the baseline codebook.
Embodiments of the invention applied to the receiver unit 110 and the transmitter unit 105 are described below using Figures 2 and 3.
Figure 2 (along with Figure 1) illustrates a flow diagram for receiver operation 200 in accordance with an embodiment of the invention. Receiver operation 200 may be descriptive of operations taking place in a receiver, such as receiver unit 110 which may be a mobile user unit. Receiver operation 200 may occur periodically as specified in a technical standard or description of a wireless communications system wherein transmitter unit 105 and receiver unit 110 operate. Alternatively, receiver operation 200 may occur each time receiver unit 110 receives a special transmission from the transmitter unit 105, or when a specified metric (such as an error rate, successful transmission rate, or so forth) meets or does not meet a specified threshold, and so forth, the receiver unit 110 may report CSI to the transmitter unit 105.
Receiver operation 200 may begin with reading the baseline codebook and the calculating the current channel correlation matrix (box 201). The current channel correlation matrix may be a time averaged information. Then, the receiver unit 110 transforms the baseline codebook using the channel correlation matrix (box 205). The baseline codebook may be any suitable codebook including Grassmannian line/subspace packing (GLP) based codebooks, DFT based codebooks, complex Hadamard transform (CHT) or Hadamard-based, Householder, Lloyd-Max algorithm, and the like.
The transforming of the baseline codebook may follow techniques described above, such as using equations (11), (14), (15), (16), or (17). In one embodiment, the receiver unit 110 transforms the baseline codebook after receiving a signal from the transmitter unit 105 instructing receiver unit 110 to transform the baseline codebook, for example. Alternatively, the receiver unit 110 transforms the baseline codebook periodically at specified time intervals (e.g., agreed upon time intervals), or when an event occurs (such as an error rate, a success rate, or some other metric meets a specified threshold), and so forth.
After transforming the baseline codebook (which also occurs at transmitter unit 105), the receiver unit 110 and the transmitter unit 105 synchronize to ensure both are using the same codebook at a given time (box 206). Then, receiver unit 110 estimates a channel, such as channel 115, over which transmitter unit 105 is transmitting information to receiver unit 110 (box 210). Receiver unit 110 may estimate channel 115 and quantize this information as CSI. The receiver unit 110 reports channel state information (CSI) back to the transmitter unit 105 by using a codeword. Based on the estimated channel/CSI, receiver unit 110 selects the codeword from the transformed codebook (box 215). For example, the receiver unit 110 selects a particular codeword to maximize the aggregate number of data bits at all of its receiving antennas. In various embodiments, the receiver unit 110 quantizes the channel using similar techniques for both SU-MIMO and MU-MIMO. The receiver unit 110 selects the index corresponding to the selected codeword.
Receiver unit 110 may then feedback the chosen index to the transmitter unit 105 over a feedback link 120 through the channel 115 (box 220). Receiver operations 200 may then terminate.
Figure 3 illustrates a flow diagram for transmitter operation 300 in accordance with an embodiment of the invention. Transmitter operation 300 may be descriptive of operations taking place in a transmitter, such as transmitter unit 105, when transmitter unit 105 receives CSI from a receiver, such as receiver unit 110. Similar to the receiver operation 200, in various embodiments, the transmitter operation 300 may occur periodically as specified in a technical standard or description of a wireless communications system wherein transmitter unit 105 and receiver unit 110 operate. Alternatively, transmitter operation 300 may occur after transmitter unit 105 receives CSI from receiver unit 110, and so forth.
Transmitter operation 300 may begin with the transmitter unit 105 reading the baseline codebook. The transmitter 110 also receives channel correlation matrix from the receiver unit 110 (box 301). The transmitter unit 105 transforms the baseline codebook (box 305) using the techniques described above with respect to equations (11), (14), (15), (16), or (17). Transmitter unit 105 may transform the baseline codebook periodically at specified time intervals, or when an event occurs (such as an error rate, a success rate, or some other metric meets a specified threshold), and so forth. The receiver unit 110 may perform the transformation of the baseline codebook at about the same time as transmitter unit 105 to ensure that both transmitter unit 105 and receiver unit 110 are using the same codebook. Further, the receiver unit 110 and the transmitter unit 105 synchronize to ensure both are using the same codebook at a given transmission (box 306).
Transmitter unit 105 may receive an index to a codeword in the transformed codebook from the receiver unit 110 (box 310). The codeword may be a quantized version of an estimated CSI and may be selected from the transformed codebook by the receiver unit 110. In general, transmitter unit 105 may receive feedback from various receivers and may use the codewords retrieved from the transformed codebook to generate a precoding matrix (box 315) and choose beamformers accordingly. In one case, for selecting beamformers, the transmitter unit 105 uses the same codeword index as received from the receiving unit 110. Alternatively, in another embodiment, the transmitted unit 105 may derive beamformers based on the codeword index. However, the transmitter unit 105 may also select the beamformers differently for SU-MIMO and MU-MIMO. For example, the transmitter unit 105 may also use information from other receivers in choosing the precoding matrix.
Based on the precoding matrix, information symbols (data to be sent) is precoded (box 320). For example, the mapping of the information symbols to each antenna is accomplished using the precoding matrix. The data thus precoded is transmitted into the channel to be received by the receiver unit 110 (box 325). Transmitter operation 300 may then be stopped.
In various embodiments, the transformation of the baseline codebook adapts the codebook to the spatial channel correlation (or other long term information) for higher rank transmissions. For each codeword, its columns are transformed differently where the transformations may be obtained from the spatial channel correlation matrix or a function thereof. In general, the transformation may be represented by matrix multiplications.
In various embodiments, a new codebook is generated, having the same size as the original baseline codebook. As the spatial channel correlation changes slowly, the transmitter updates its knowledge of transmitter side channel correlation matrix R based on an infrequent feedback or uplink sounding, and accordingly updates the codebook as well. Thus, a common codebook is agreed upon between the transmitter and receiver at this point.
Although the present invention and its advantages have been described in detail, it should be understood that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims. For example, many of the features and functions discussed above can be implemented in software, hardware, or firmware, or a combination thereof
Moreover, the scope of the present application is not intended to be limited to the particular embodiments of the process, machine, manufacture, composition of matter, means, methods and steps described in the specification.

Claims

A method of wireless communication comprising:
at a processor, reading a baseline codebook of at least a rank-2, the baseline codebook comprising codewords representing precoding matrices; and

generating an adaptive codebook by:
multiplying a first column of a codeword of the baseline codebook with a first transform matrix calculated from a channel correlation matrix, and

multiplying a second column of the codeword with a second transform matrix calculated from the channel correlation matrix, wherein the first and the second transform matrices are orthogonal.
The method of claim 1, wherein the baseline codebook is a rank-2 codebook, and wherein the first transform matrix comprises a summation using the odd eigenmodes of the channel correlation matrix, and wherein the second transform matrix comprises a summation using the even eigenmodes of the channel correlation matrix.
The method of claim 1, further comprising:
after generating an adaptive codebook, at a receiver, selecting a codeword index of the adaptive codebook to quantize channel state information;

transmitting the selected codeword index to a transmitter; and

at the transmitter, selecting beamformers for transmitting data to the receiver based on the selected codeword index.
The method of claim 1, wherein the adaptive codebook is generated at both a receiver and a transmitter, and wherein the transmitter and the receiver exchange information to synchronize the use of the adaptive codebook.
The method of claim 4, further comprising:
calculating a channel correlation matrix before generating an adaptive codebook;

after exchanging information to synchronize, receiving a codeword index from a receiver;

retrieving a codeword from the adaptive codebook;

generating a precoding matrix from the codeword; and

precoding data to be transmitted using the precoding matrix.
The method of claim 1, wherein the adaptive codebook is generated periodically at specified time intervals; or
the adaptive codebook is generated when a predefined criterion is met, wherein the criterion comprises a metric requiring a predefined threshold.
The method of claim 1, wherein generating an adaptive codebook further comprises multiplying a third column of the codeword with a third transform matrix calculated from the channel correlation matrix, wherein the third transform matrix is orthogonal with the first and the second transform matrices.
The method of claim 7, wherein the baseline codebook is a rank-3 codebook, and wherein the first transform matrix comprises a summation using 3k-2 eigenmodes of the channel correlation matrix, wherein the second transform matrix comprises a summation using 3k-1 eigenmodes of the channel correlation matrix, wherein the third transform matrix comprises a summation using 3k eigenmodes of the channel correlation matrix, and wherein k is a natural number less than the total number of transmitting antennas in a transmitter.
A method of wireless communication comprising:
at a mobile user device, generating an adaptive codebook by transforming a predefined baseline codebook of at least a rank-2, the baseline codebook comprising codewords representing precoding matrices, wherein the transforming comprises:
multiplying a first column of a codeword of the baseline codebook with a first transform matrix calculated from a channel correlation matrix, and

multiplying a second column of the codeword with a second transform matrix calculated from the channel correlation matrix, wherein the first and the second transform matrices are orthogonal.
The method of claim 9, further comprising, using the adaptive codebook, transmitting to a transmitter, a quantized form of a channel between the mobile user device and the transmitter.
The method of claim 9, wherein generating an adaptive codebook further comprises multiplying a third column of the codeword with a third transform matrix calculated from the channel correlation matrix, wherein the third transform matrix is orthogonal with the first and the second transform matrices.
The method of claim 11, wherein the baseline codebook is a rank-3 codebook, and wherein the first transform matrix comprises a summation using 3k-2 eigenmodes of the channel correlation matrix, wherein the second transform matrix comprises a summation using 3k-1 eigenmodes of the channel correlation matrix, and wherein the third transform matrix comprises a summation using 3k eigenmodes of the channel correlation matrix, wherein k is a natural number less than total number of transmitting antennas in a transmitter.
A method of wireless communication comprising:
at a base station, generating an adaptive codebook by transforming a predefined baseline codebook of at least a rank-2, the predefined baseline codebook comprising codewords representing precoding matrices, wherein the transforming comprises:
multiplying a first column of a codeword of the predefined baseline codebook with a first transform matrix calculated from a channel correlation matrix, and

multiplying a second column of the codeword with a second transform matrix calculated from the channel correlation matrix, wherein the first and the second transform matrices are orthogonal.
The method of claim 13, further comprising:
at the base station, receiving a codeword index from a mobile user device;

retrieving a codeword by looking up the index in the adaptive codebook;

generating a precoding matrix from the codeword; and

precoding data to be transmitted using the precoding matrix.
The method of claim 13, wherein generating an adaptive codebook further comprises multiplying a third column of the codeword with a third transform matrix calculated from the channel correlation matrix, wherein the third transform matrix is orthogonal with the first and the second transform matrices.
The method of claim 13, wherein the baseline codebook is a rank-3 codebook, and wherein the first transform matrix comprises a summation using 3k-2 eigenmodes of the channel correlation matrix, wherein the second transform matrix comprises a summation using 3k-1 eigenmodes of the channel correlation matrix, and wherein the third transform matrix comprises a summation using 3k eigenmodes of the channel correlation matrix, wherein k is a natural number less than the total number of transmitting antennas in the transmitter.
A mobile user device for wireless communication adapted for using the method according to any of claims 9 to 12.
A base station for wireless communication adapted for using the method according to any of claims 13 to 16.